Tag
Elorian AI 发布 Glimpse 模型,声称是首个实现「视觉思考」的 AI 模型,能够像人类一样逐步观察、检查并推理图像内容,在 12 个未见过的迷宫测试中成功解出 11 个,并自行编写代码追踪路径并绘制动画解答。
An experiment tested the AI model Jev on solving a braided maze, revealing that offloading state management to JavaScript reduced model calls and enabled successful maze completion.
Explores whether large language models can solve maze navigation tasks, probing their reasoning and spatial understanding abilities.